The sample consists of Gorilla gorilla (n= 27), Pan troglodytes (n= 16), Homo sapiens
(n=44) from The Field Museum in Chicago, and Australopithecus africanus (n=24), and
Australopithecus robustus (n=9) (Table 1) dental measurements from Dr. Frank Williams’ research in South Africa (Table 2). Each of the individuals from The Field Museum are
associated with sex and age estimations. The measurements obtained during the trip to The Field Museum have been utilized to test if there is a dental size difference among StW 252,
Australopithecus africanus, and Australopithecus robustus with a much larger more diverse comparative sample.
Table 1. Sample size and location
Taxa Number of Individuals Location
Gorilla gorilla 27 Chicago Field Museum
Pan troglodytes 16 Chicago Field Museum
Homo sapiens 44 Chicago Field Museum and Georgia State University
Australopithecus africanus 24 University of Witwatersrand and Transvaal Museum
Australopithecus robustus 9 Transvaal Museum
Table 2. A. africanus and A. robustus measurements used in study
Taxa Specimen # A. africanus StW 53, MLD 9, MLD 6, MLD 45, MLD 28, MLD 11, StW 40, StW 73, StW 404, StW 14, StW 183, StW 138, StW 126, StW 127, StW 188, StW 287, StW 189, StW 133, StW 140, StW 132, StW 131, StW 56, TM 1511, TM 1512 A. robustus Sk 46, Sk 48, Sk 52, Sk 83, Sk 47, Sk, 13/14, Sk 11, Sts 52, TM 1517 4.1.1 Measurement Methods
Seven modern humans from the Georgia State University skeletal collection were measured to collect preliminary sample dental measurements. These individuals are included in
the larger comparative sample increasing the number of H. sapiens individuals to 44 (Table 1). Dental measurements were acquired from both left and right side of the modern human sample to document a full set of measurements. The dental measurements consisted of buccolingual and mesiodistal measurements of the maxillary teeth using digital calipers. The maxillary teeth were measured because these are the only teeth present for StW 252. Each tooth was individually measured three times beginning with the buccolingual angle and then the mesiodistal angle on incisor1, incisor2, canine, premolar3, premolar4, molar1, molar2, and molar3. The fossil hominin sample used for this study is a collection of unpublished dental measurements collected by Dr. Frank Williams from his research in South Africa (Table 2). His field journal contains
measurements of both mandible and immature dentition, but these measurements were omitted since there is no mandible found in association with StW 252 and the maxillary dentition is nearly fully erupted. This process was repeated during data collection for the larger comparative sample at The Field Museum Chicago. However, each individual was only measured once instead of three times.
.
4.1.2 Measurement Error Analysis
To control for measurement error, each individual was measured in three trials. Each trial was recorded in an Excel spreadsheet and then transferred to SPSS to test for inter-observer error. The anticipated measurement error was less than 0.5 mm. To calculate the measurement error, each of the three trials was averaged. To estimate measurement error, the mean for each dental measurement must be calculated and compared to each of the three trials using the absolute value (ABS). The equation used to calculate the mean of each tooth is:
“X” is the average of the 3 trials. For specimen GSU 57.1 which totaled 7.53 mm, the equation is now:
ABS (7.9-7.53) + ABS (7.38-7.53) + ABS (7.3- 7.53)/3
The values for each individual were then compared using a one-way ANOVA test to determine if the variation between trials exceeded the variation between individuals (Appendix 1; Table 9 and 10). The p-value of the ANOVA is 0.05.
4.1.2.1 Measurement Error Analysis Results
The significance value of each ANOVA is < 0.900 indicating no real groups. In other words, the differences between trials was less than the differences between individuals. To further estimate measurement error, the minimum and maximum absolute mean deviations were
identified. The highest mean deviation is the 3rd molar buccolingual measurement at 3.84 mm
and the highest absolute mean deviation is 0.95 mm. The lowest absolute mean deviation is 0.23 mm for the fourth premolar with a minimum deviation of 0.04 mm and a maximum deviation at 0.59 mm. For each tooth measurement, the average measurement error was calculated by taking the average of the deviations of each of the three measurements from the mean of all three measurements (Table 4). The average of all of the averages was found by taking 16 numbers in Table 3 and adding them together followed by dividing the numbers by the number of variables (16). As mentioned previously, the absolute mean measurement error is 0.3853 mm. This measurement fails to show a meaningful difference between each measurement. The maximum amount of deviation between future measurements is anticipated to be less than 0.3853 mm. It is expected that measurement precision will improve as this study progresses.
4.2 Analytical Methods
The analytical analysis includes a univariate sex specific analysis of each measurement angle for each tooth except for the three molars for which the geometric mean was utilized. The bivariate scatterplots were completed to determine to which group StW 252 most closely align using 95% confidence ellipses around group centroids to demarcate taxa. This method was
modified from a study conducted on RH1, a single I2, done by Kramer et al. (2005). For this
project, the variables tested are the buccolingual (BL) and mesiodistal (MD) angles for each dental measurement. Eight charts were completed for each dental measurement set. The charts
were separated by MD (y axis) and BL (x axis) measurements for I1, I2, C, P3, P4, M1, M2, and
M3 for each specimen. An ANOVA was conducted to determine if the groups differed for each
dental measurement. Additionally, a Tukey’s Post-hoc test was executed to see if there exists a pair-wise difference among groups. A principal components analysis using the five dental
dimensions most often preserved in A. africanus and A. robustus was calculated and the first two
PC axes plotted using 95% confidence ellipses around group centroids to identify where StW
252 would fall with respect to the comparative taxa. The discriminant function analysis coupled
with the Mahalanobis’ Distances and Jackknifed and Non-Jackknifed classification rates includes only the molar measurements scaled to the geometric mean to classify StW 252 with the
comparative sample. The p-value for these tests are 0.05. The means for each taxon were taken
for the cluster analysis using all dental traits except I2 BL as this was not preserved for A.
robustus. The data were scaled to the geometric mean of all 15 dental measurements before the cluster analysis was conducted to remove the effects of size. Lastly, a qualitative analysis of StW 252 and Sts 71 craniofacial morphology was conducted.
5 RESULTS